APRIL 15, 1937
ANALYTICAL EDITION
give a more reliable figure than t o take the average of ten readings at any one wave length. The spectrophotometric method, provided a spectrophotometer possessing the precision of the Purdue instrument is used, is capable of giving results for iron in ores which will be within k0.30 per cent of the values given by the dichromate titration method and many will be within kO.10 per cent. Results may be duplicated on the same sample with a precisionof about = t O . l O to +0.20 per cent. TABLE11. COMPARISON OF PERCENTAGE ERRORS IN IRONAND COPPER DETERMINATIONS
--
Iron, 12 samples Copper, 11 samples
Percentage ErrorMaximum Minimum 0.65 0.00 1.98 0 00
Average 0.20 0 69
In Table I1 is shown a comparison of the reliability of the results for iron, on a percentage basis, with those previously obtained by the writer (4) for copper ores, where the actual percentages were much lower. The figures indicate that the 'spectrophotometric method involving the use of the molecular extinction coefficient can be satisfactorily extended to the determination of iron where the percentage of the constituent determined is relatively high.
Summary The spectrophotometric method employing the molecular extinction eoefficient has been applied t o the determination
163
of iron in ores. Although the iron content is relatively high, the percentage error is smaller than when similar transmittancy measurements were made on the copper-ammonia system where the copper content was relatively very low. The spectrophotometric method gives results for iron in ores which check well with those given by the dichromate method.
Acknowledgments The writer wishes t o express his sincere appreciation to 111.G. Mellon, Department of Chemistry, Purdue University, in whose laboratory this investigation was conducted, and to thank him for his kind interest, helpful criticism, and profitable suggestions and for the use of the Purdue spectrophotometer. Thanks are also given to H. W. Swank, G. Dragt, and E. R. Wright of Purdue for their aid in adjusting the spectrophotometer.
Literature Cited (1) Beoh, P. F., Dan& Tids.Farm., 9 , 2 8 9 (1935). (2) Mahin, E. G., "Quantitative Analysis," p. 242, New York, McGraw-Hill Book Co.. 1932. (3) Mehlig, J. P., IND.Eaa. CHEM.,Anal. Ed., 7, 2 7 (1935). (4) Ibid., 7 , 3 8 7 (1935).
(5) Snell, F. D., and C. T., "Colorimetric Methods of Analysis," Vol. I, p. 301, New York, D. Van Nostrand Co., 1936.
RECEIVED September 1, 1936.
Determination of Acetylmethylcarbinol Effect on Certain Analytical Procedures A. F. LANGLYHKE
S
AND W.
H. PETERSON, University of Wisconsin, Madison, Wis.
INCE acetylmethylcarbinol is found in varying quantities among the fermentation products of a large variety of bacterial cultures and enzymatic preparations (3, 8, 11, 13, 14, 15), a knowledge of its characteristics and its effect on various analytical procedures is particularly desirable. Because of the nature of the compound it reacts in many of the common analytical procedures, causing error in the final results. Thus, because of its reducing properties it interferes in the common methods for the determination of reducing sugars. Since it is volatile it appears in distillates and causes error in the analyses for volatile products. It reacts with alkaline iodine and therefore interferes with the ordinary method for the determination of acetone (6). It gives rise to acid products on oxidation and in this way interferes with methods for the determination of alcohol which depend on oxidation. The acetylmethylcarbinol used for these experiments was obtained from the Lucid01 Corporation, Buffalo, N. Y. The liquid material was crystallized by holding for one month a t a temperature of approximately -10" C. The material so obtained was of a mushy consistency and showed a yellow coloration and pungent odor which were removed by washing several. times with ether. The resulting white crystalline powder gave a melting point which varied from 84" to 125"C., depending on the rate at which the oil bath was heated. Statements in the literature indicate that the melting point of dimeric acetylmethylcarbinol is rather irregular and varies with the method of crystallization ( I , 2). Determination of the molecular weight of the crystalline material by elevation of the boiling point indicated an initial value of about 160 which dropped to 101 after boiling for 25
minutes because of conversion to the monomer. Other investigators (I,,%') have found molecular weights for acetylmethylcarbinol dimers which ranged from 170 t o 195 as compared to the theoretical value of 176.
Reaction with Alkaline Iodine Since acetylmethylcarbinol contains the acetyl group (CHaCO-) directly linked to carbon, it will react with alkaline iodine to form iodoform (5). To verify the nature of the reaction between acetylmethylcarbinol and iodine, aliquots of standard aqueous solutions of the compound (containing 3 to 10 mg.) were made alkaline with 1 N sodium hydroxide and standard iodine solution was added, Iodoform was precipitated and, after standing for 10 to 15minutes, the excess iodine was liberated by the addition of 1N sulfuric acid and titrated. The averages of a large number of determinations are presented in Table I. The iodine used approaches six atoms per molecule of iodoform, which supports the following formulation for the reaction: 31a 6NaOH = 3NaOI 3NaI 3Hz0 CHsCHOHCOCHs 3NaOI = CHI* CHsCHOHCOONa 2NaOH
+
+
+
+
+
+
The iodoform isolated from one reaction sublimed a t 118" to 119" C. as compared to the reported value of 119" (9). Lactic acid was recovered from the products of the reaction by extraction of the acidified solution with ether. From the ether extract the zinc salt was prepared and analyzed. The water of crystallization was determined by drying to a constant weight at 110" C., the zinc content by ignition, and lactic
164
INDUSTRIAL AND ENGINEERING CHEMISTRY
OF ACETYLMETHYLCARBINOL WITH ALKATABLEI. REACTION LINE IODINE
0.1 N IZ per Mg. CHaCHOHCOCHa
Sample
Atoms I per Molecule CHsCHOH- Per Cent of COCHs Theory
cc.
Crystals washed with ether Ether-washed product recrystallized from acetone Theoretical for reaction
0.671
5.91
98.5
0.672
5.92
98.7
0.681
6.00
100.0
(k (k
= 1.239 to 1.245) = 1.940 to 2.035).
25 50 75
-Acetylmethylcarbinol in DistillateConcentration of Solution: 0.005%.- 0 . 0 2 % 0.072% 0.10% ._ ..
Av.
%
%
%
%
%
31.8 59.7 83.7
32.1 60.3 83.9
31.5 59.6 83.1
5912
31.8 59.7 83.6
..
acid by the method of Friedemann and Graeser (4) with the following results : Calcd. for Zn(C3H60a)~3H20:Hz0 18.2. Found: 18.0 Calcd. for Zn(C3HsOs):Zn 26.9, CH3CHOHCOOH74.0. Found: Zn 29.0, CHaCHOHCOOH 73.9.
and less volatile than butyric acid
I n the determination of acetone in a fermented medium containing acetylmethylcarbinol, a correction must be made which depends on the concentration of acetylmethylcarbinol and the volume of distillate collected. If 25 per cent of the liquid is distilled off, approximately 32 per cent of the acetylmethylcarbinol will be contained in the distillate. For other volumes distilled, the correction may be applied most conveniently by plotting a curve using the values of Table 11,or the correction may be made by calculation using the equation 1 log E! = k log 2 -
TABLE11. VOLA~ILITY OF ACETYLMETRYLCARBINOL IN AQUEOUS SOLUTION Fractions, Percentage of Total Distilled
VOL. 9, NO. 4
Yz
22
When 35 per cent of the liquid is distilled, ??= 2 and setxz 65' ting k equal to 1.31 the value for y2, the amount of acetyl-
4.0
roz5I
I A
I
The data show that the product isolated was nearly pure inaotive zinc lactate containing three molecules of water of crystallization. The high zinc content is probably due to the presence of small amounts of excess zinc-e. g., zinc hydroxidewhich raise the zinc value but have no appreciable effect on the other analyses.
Volatility The rate of distillation of acetylmethylcarbinol from aqueous solution was determined by a modification of the method of Virtanen and Pulkki (18). I n order to prevent condensation and reflux, an insulating hood made from a large metal can was used to cover the distillation flask and in order to ensure uniform heating, a metal shield was used to protect the Bunsen burner. I n the experiments 200 cc. of solution were distilled and three 50-cc. fractions were collected. These fractions as well as the residue were analyzed by application of the iodoform reaction. The percentage of acetylmethylcarbinol in the distillates was found to be independent of the rate of distillation when the precautions to prevent reflux were observed. The data for the distillation of solutions varying in concentration from 0.005 to 0.1 per cent are presented in Table 11. As shown by the data, the distribution of acetylmethylcarbinol during distillation is constant and independent of the concentration of the solution. A quantitative expression for the volatility may be calculated by application of the formula used by Virtanen and Pulkki (18): k = -log ?/I - log Y2 log 2, - log 22
I n this equation yl represents the amount of volatile compound in the solution at the beginning of distillation and y2 the amount a t the end, while 21 represents the amount of water a t the beginning and 22 the amount at the end. The values for the distillation constant, k , as calculated from the average figures for the first, second, and third fractions (Table 11) were 1.330, 1.297, and 1.298. On comparison with the figures of Virtanen and Pulkki (18) it is seen that acetylmethylcarbinol is a little more volatile than propionic acid
0.5
I O MILLIGRAMS
ACETYLMETHYL
I.'5
CARBINOL
1. CUPROUS OXIDE PRODUCED I N REDUCTION OF SHAFFER-HARTMANN SUGAR REAffENTBY ACETYLMETHYLCARBINOL
FIffURE
methylcarbinol remaining undistilled, is found to be 56.8 per cent. Therefore, 43.2 per cent of the acetylmethylcarbinol will be found in the distillate.
Reaction with Oxidizing Reagents I n order to determine the corrections to be applied in the analyses for reducing sugar, two methods for the determination of glucose were investigated. ,These were the ShafferHartmann method as modified by Stiles, Peterson, and Fred (17),which depends on the reduction of copper and iodometric determination of the cuprous oxide produced, and the Hagedorn-Jensen method (7), which depends on the reduction of potassium ferricyanide to ferrocyanide and titration of t h e excess ferricyanide with dilute standard thiosulfate solution. In each case varying amounts of an aqueous solution of acetylmethylcarbinol were submitted to the procedure used in the determination of the reducing sugars and the extent of reduction of the reagent was determined. The results are shown graphically in Figures 1 and 2. REDUCTION OF CUPRICSULFATE.According to Kling (12), acetylmethylcarbinol is oxidized by alkaline cupric oxide
ANALYTICAL EDITION
APRIL 15,1937
to acetic acid, but such a result was not obtained in these experiments. Stahly and Werkman (26) report that 3.01 grams of cuprous oxide are produced from each gram of acetylmethylcarbinol under the conditions of the MunsonWalker method. I n the copper-reduction method used, the tubes are only loosely stoppered to prevent oxidation by the air and loss of volatile products is not prevented. It is also possible that the 15-minute reaction period is not long enough for complete reaction. For these reasons the amount of copper reduced does not bear a stoichiometric relation to the amount of acetylmethylcarbinol in the sample, although there is a regular relation between these two variables. If acetylmethylcarbinol is oxidized to diacetyl, two hydrogen equivalents; per mole are required, while for the oxidation to acetic acid four hydrogen equivalents are necessary. As calculated from the graph of Figure 1, 2.95 hydrogen equivalents are actually used in the oxidation, which shows that only about one-half of the diacetyl formed is further oxidized to acetic acid. The points described a straight line, which does not, however, pass through the origin. This apparent anomaly can be explained only on the assumption that a constant small amount of acetylmethylcarbinol escapes oxidation, or that a constant small amount of diacetyl is lost through evaporation. REDUCTION OF FERRICYANIDE. As in the reduction of copper, the reduction of ferricyanide proceeds without complete conversion of the acetylmethylcar'3inol to acetic acid. I n this case 2.67 hydrogen equivalents are required, although more variation was observed in this reaction. This is shown by the fact that the individual points on the graph of Figure 2 do not correspond so well to a straight line as in the case of the reduction of copper.
Oxidation with Acid Potassium Dichromate I n the determination of butyl and ethyl alcohols by the method of Johnson (IO),a distillate containing the neutral volatile products from a bacterial fermentation is oxidized by means of acid potassium dichromate solution and the acid resulting from the oxidation is removed by distillation. Two fractions are collected and the content of butyl and ethyl alcohols in the original sample is calculated from the titrations of these fractions. Since acetylmethylcarbinol yields
ACETYLMETHYL CARBINOL
FIGURE 2. REDUCTION OF POTASSIUM FERRICYANIDE
(HAQEDORN-JENSEN
REAQENT)
ACETYLMETHYLCARBINOL
BY
165
acid products on oxidation, it is of interest to know what effect this compound will exert in the alcohol method. TABLE111. EFFECT ON JOHNSONMETHODFOR ETHYL ALCOHOLS 7
Acetylmethylcarbinol Sample
Butyl alcohol
Mo.
MO
Bctual
.
5 -0.01 10 0.14 15 0.14 Av. (pef mg. acetoin) 0.007 Based on assumption that one lent t o two moles of ethyl alcohol.
BUTYLAND
.4lcohol Indicated Ethyl alcohol
Theoretical'' Butyl Ethyl alcohol alcohol
Mu.
Mo.
Mo.
5.26 10.48 15.87
0
6.23 10.46 16.69
0 0
1.053 0 1.046 mole of acetylmethylcsrbinol is equiva-
Samples of 5, 10, and 15 mg. of acetylmethylcarbinol in aqueous solution were analyzed according to the procedure for the determination of alcohols by the Johnson method (IO). In Table I11 the results obtained are expressed as milligrams of butyl and ethyl alcohols. The oxidation of acetylmethylcarbinol by the acid dichromate yields acetic acid, and therefore for purposes of correcting any alcohol determinations one mole of acetylmethylcarbinol corresponds to two moles of ethyl alcohol.
Determination of Acetylmethylcarbinol Since acetylmethylcarbinol reacts quantitatively with alkaline iodine reagent and since the carbinol is distilled from aqueous solution a t a definite rate which is independent of the concentration, it was possible to develop a method for its determination in which these properties were applied. Attempts to separate the carbinol from the fermented medium were unsuccessful, as neither ether extraction nor constant volume distillation gave a quantitative recovery of the compound. I n order to apply the iodoform reaction directly to the determination of acetylmethylcarbinol it is necessary to eliminate from the sample any other compounds which will react with the iodine reagent. Acetone is commonly present in large excess over the acetylmethylcarbinol content in some fermentations, but this compound is practically completely eliminated by distillation of half the volume of an aqueous solution. Ethyl alcohol in appreciable concentration will also react with alkaline iodine, although this reaction is very incomplete at room temperature and in the short reaction time used. However, it is also completely eliminated by half distillation. Butylene glycol is also likely t o occur in association with acetylmethylcarbinol in fermentations and will react with alkaline iodine in 15 minutes at room temperature to produce an apparent acetylmethylcarbinol content equivalent to 10 per cent of its own weight. However, because of the nonvolatility of this compound, the error resulting by analysis of the third quarter of distillate is negligible. By far the most convenient method for the determination of acetylmethylcarbinol involves direct distillation from the fermentation mixture, and by this method quantitative results could be obtained. The disadvantage of direct distillation from the culture medium lies in the fact that many of the media are prone to foam badly during distillation. However, many ordinary media may be directly distilled in the fractional distillation procedure and certain procedures may be applied to minimize foaming when this characteristic is displayed. It was found that a mineral salts-peptone-glucose medium which had been fermented by A . aerogenes in the presence of an excess of calcium carbonate could not be distilled, because of excessive foaming. To correct this condition the culture was acidified with sulfuric acid, heated just to boiling and, after cooling, filtered by auction through an asbestos mat.
INDUSTRIAL AND ENGINEERING CHEMISTRY
166
VOL. 9, NO. 4
The results reported in Table IV were obtained on a variety TABLEIV. DIRECTDETERMINATION OF ACETYLMETHYLCARBI- of cultures as shown. The salts-peptone medium used conNOL BY FRACTIONAL DISTILLATION OF CULTURE sisted of 0.07 per cent dibasic ammonium phosphate] 0.5 per Acetylmethylcarbinol cent peptone, 0.1 per cent asparagine, and tap water. In the Found Recovery of Added (after addiAcetylmethylcase of the fermentations 3 per cent of glucose was added. The Sample Found Added tion) carbinol corn mash was made up with 6 per cent of corn meal in tap Mg./100 !Mg./100 Mg./lOO Mg.//OO water, In the second column of Table IV are given the concc. CC. cc. % 49.2 50.3 50.3 102.1 centrations of acetylmethylcarbinol as determined directly on Sterile saltsthe culture by the procedure outlined. Acetylmethylcarbinol peptone medium . 21.0 21.4 21.4 101.9 CZ. polyrnyxa in was then added in the amounts given in the third column and 6% corn mash 40.3 49.6 89.5 49 2 99.2 the content of the samples so prepared was again determined A. aerogenes in salts-peptone with the results expressed in the fourth column. I n the medium 29.1 37.8 67.9 38.8 102.7 fifth and sixth columns are entered the recoveries of the added A . aeroaenes in salts-peptone acetylmethylcarbinol as obtained from the difference between medium (excess the determinations before and after the addition of a known CaCOs) 39.4 44.0 85.1 45.7 103.9 A. aerogenes in quantity of acetylmethylcarbinol. The data reported are the salts-peptone averages of triplicate analyses. The deviation of individual 18.2 10.0 28.1 9.9 99.0 medium Cl. ecetobutyliresults from the average ranged from 0 to 4.1 per cent and cum in 6% averaged 1.13 per cent. The average error in the recovery corn mash 37.4 49.8 87.1 49.7 99.8 of added acetylmethylcarbinol was 2.2 per cent. The recoveries are in general a little high, and the average recovery is about 101.2 per cent. Therefore] a somewhat The resulting product showed considerable less tendency to higher degree of accuracy may be obtained if the recovery is foam, but as an added precaution the sample was diluted to determined separately for each different medium used, as four times the volume before applying the distillation prothis value will vary somewhat with the nature of the solution cedure. By this procedure quantitative results could be under analysis. Thus, if instead of using the value of 23.9 obtained. per cent recovery as determined for aqueous solutions a corFor the results reported in Table IV the following general rected value of 23.6 per cent is used, the errors in the deprocedure was used: terminations reported would vary from -2.2 to $2.7 per Enough of the culture to give 100 cc. of clear filtrate is filtered cent. through a fluted filter and 100 cc. of the clear filtrate are then pipetted into the 300-c~.distillation flask which cantains 6 or 8 Summary glass beads to revent bumping. The flask is directly connected to a short contenser by means of a ground-glass joint and is supBecause of the unique nature of the compound, acetylorted by a rin bearing a wire gauze with an asbestos mat center. %he insulatingfoodis placed over the flask, and heating is started methylcarbinol affects a variety of common analytical prowith a burner whichis protected by a shield so as to guarantee unicedures. Methods for the correction of acetone, alcohol, form heating. The rate of heating is so adjusted that the distiland reducing sugar determinations in the presence of acetyllate which collects is cold and the medium under distillation boils methylcarbinol are described. Based on its volatility and evenly. A fraction of exactly 50 cc. is collected in a volumetric flask and may either be discarded or used in the determination iodoform reaction, a method for its quantitative determinaof the more volatile products of the fermentation. A second tion has been developed. fraction of exactly 25 cc. is then collected and analyzed by the iodoform reaction. The contents of the 25-cc. volumetric flask are added to 15 CC. Literature Cited of 1N sodium hvdroxide in a 150-cc. Erlenmever flask. The volumetric flask 'Is rinsed with 3 portions of a few cubic centi(1) Diels, O., and Stephan, E., Ber., 40, 4336 (1907). meters of water and the rinsin s are added t o the sample in the (2) Dirscherl, W.,and Braun, E., Ibid., 63B,416 (1930). Erlenmeyer. Then while the flaak is constantly shaken in one (3) Elion, L., Biochem. Z., 171, 40 (1926). hand, exactly 5 cc. of 0.2 N iodine are added by rapid drop ing (4) Friedemann, T.E., and Graeser, J. B., J . Biol. C h m . , 100,291 from a ipet held in the other. The precipitation of iodogrm (1933). begins a?most immediately and after 10 minutes the reaction is ( 5 ) Fuson, R. C., and Bull, B. A., Chem. Reo., 15, 276 (1934). complete. After standing for 10 to 15 minutes out of direct (6) Goodwin, L.F.,J . Am. Chem. Boo., 42, 39 (1920). sunlight 20 cc. of 1 N sulfuric acid are run into the flask and the (7) Hagedorn, H. C., and Jensen, B. N., Biochm. Z., 135,46 (1923). liberated iodine is titrated immediately to the starch end point (8) Horovitz-Vlassova, L. M.,and Rodionova, E. A,, Zentr. Bakt. Parasitenk., I1 Abt., 87, 333 (1933). with standard 0.1 N sodium thiosulfate solution. A semimicroburet which may be read accurately to 0.01 cc. is used. A blank (9) Intern. Crit. Tables, Vol. I, p. 176, New York, McCraw-Hill for the purpose of standardizing the iodine solution is run on Book Co.,1926. 5 cc. of the iodine solution, with distilled water replacing the (10) Johnson, M.J., IND.ENQ.CHEX.,Anal. Ed., 4, 20 (1932). (11) Kay, H. D., Biochm. J.,20, 321 (1926). sample in the ordinary procedure.
1. "
Since one mole of acetylmethylcarbinol is equivalent to 6 atoms of iodine, one millimole (or 88 mg.) is equivalent to 60 cc. of 0.1 N iodine. Then, for each cc. of 0.1 N iodine used in the iodoform reaction there are 1.467 mg. of acetylmethylcarbinol in the sample. According to Table 11,23.9 per cent of the acetylmethylcarbinol in the sample appears in the third quarter of distillate. Therefore, to obtain the total acetylmethylcarbinol content of the 100-cc. sample distilled, the content of the third quarter of distillate is divided by the factor 0.239.
(12) Kling, A., Bull. aoc. chim., ( 3 ) 35, 209 (1909). (13) May6, P.,Compt. rend. soc. biol., 81, 636-7 (1918). (14) Neuberg, C.,and Reinfurth, E., Biochem. Z., 143, 653 (1923). (15) Pederson, C. S.,and Breed, R. S., J. Bact., 16, 163 (1928). (16) Stahly, G.L.,and Werkman, C. H., Iowa State CoZZ. S. Sci., 10, 205 (1936). (17) Stiles, H. R., Peterson, W. H., and Fred, E. B., J. Bact., 12, 427 (1926). (18) Virtanen, A. I., and Pulkki, L., J. Am. Cham. SOC.,50, 3138 (1928). RECEIVED October 26, 1936. Supported in part by a grant from the Special Heeearoh Fund of the Graduate School, University of Wisconsin.
